4302
Inorg. Chem. 2001, 40, 4302-4309
IR and EXAFS Spectroscopic Studies of Glyphosate Protonation and Copper(II) Complexes of Glyphosate in Aqueous Solution Julia Sheals,*,† Per Persson,† and Britt Hedman‡ Department of Chemistry, Inorganic Chemistry, Umeå University, S-901 87 Umeå, Sweden, and Stanford Synchrotron Radiation Laboratory, SLAC, Stanford University, Stanford, California 94309 ReceiVed July 28, 2000 The varying degrees of protonation of N-(phosphonomethyl)glycine (PMG, glyphosate) were investigated with infrared (IR) spectroscopy and ab initio frequency calculations. The zwitterionic nature of PMG in solution was confirmed, and intramolecular hydrogen bonding was identified. Successive protonation of the PMG molecule follows the order amine, phosphonate, carboxylate. Intramolecular hydrogen bonding is indicated to exist at all stages of protonation: between both RCO2- and RNH2+ and RPO32- and RNH2+ in HL2- (where L represents the ligand PMG); between RCO2- and RNH2+ in H2L-; predominantly between RPO32- and RNH2+ in H3L. There are strong indications that the zwitterion is intact throughout the pH range investigated. Results from IR and extended X-ray absorption fine structure (EXAFS) spectroscopies provide new evidence for structures of N-(phosphonomethyl)glycinecopper(II) complexes. The structures of 1:1 complexes, CuL- and CuHL, are essentially the same, differing only in protonation of the phosphonate group. Copper(II) lies at the center of a Jahn-Teller distorted octahedron with all three donor groups (amine, carboxylate, phosphonate) of PMG chelating with copper(II) to form two five-membered chelate rings oriented in the equatorial plane. EXAFS indicates that oxygen (most likely a water molecule) is a fourth ligand, which would thus occupy the fourth corner in the equatorial plane of the elongated octahedron. CuL24- most probably forms an isomeric mixture in solution, and there are indications that this mixture is dominated by complexes where two PMG ligands are bound to copper(II) via equatorial and axial positions, with both phosphonate and carboxylate donor groups responsible for chelation at axial positions.
Introduction The monitoring and control of pollutants in the aquatic environment is becoming of increasing concern. Detailed investigations of interactions between both naturally occurring and anthropogenic species found in the environment provide information regarding the chemical structure and speciation, and thus the transport, bioavailability, and potential toxicity of pollutants. To fully understand the chemical processes that occur in groundwater, soils, and surface waters, it is important to examine the role of the solid-water interface as well as interactions in the aqueous phase. The present paper is the first in a series of papers investigating ternary surface complexes formed during the adsorption of copper(II) and an organic ligand, N-(phosphonomethyl)glycine (PMG), onto mineral oxide surfaces. As it is beneficial to first study the adsorbates in solution before attempting to determine their behavior at the surface, only the solution chemistry of copper(II) and PMG is described here. PMG, also known as glyphosate, is a component of organophosphorus herbicides widely used in agriculture. PMG is capable of killing weeds by inhibiting the shikimic acid pathway, a biochemical pathway present in plants but not in animals. It is applied to kill crop weeds while the crops themselves are engineered to withstand exposure to the PMG.1 It is known to be immobilized in soils2 and inactivated by microorganisms to form nonphytotoxic * Author to whom correspondence should be addressed. † Umeå University. ‡ Stanford University. (1) Greenpeace ReportsNot Ready for Round-Up Fact Sheet. Greenpeace, 1997. (2) Glass R. L. J. Agric. Food Chem. 1987, 35, 497-500.
products.3 There are several reports in the literature discussing the potential toxicity of PMG.4,5 Figure 1 represents the speciation of PMG, which forms a pH-dependent zwitterion in solution. It should be noted that the PMG molecule possesses three donor groups, an amine group, a carboxylate group, and a phosphonate group, which are responsible for complexation reactions with trace metal ions and mineral oxide surfaces. Some trace metal ions, although essential to life, are harmful at excessive concentrations,6 and chelating agents such as PMG can potentially be used to reduce the bioavailable concentration of trace metals in aquatic and soil environments. Copper(II) provides an ideal example of a trace metal ion, arising from both natural and anthropogenic sources. The interactions of copper(II) and PMG, with each other and with mineral oxides, and how these affect the bioavailabilty of each of these substances are just an example of the numerous interactions occurring in surface and groundwaters. Attempting to fully understand such chemical processes in the environment requires a holistic approach, combining results from both molecular-level and macroscopic investigations. The results of this and similar molecular-level studies allow constraints to be applied in the modeling of thermodynamic data. Such chemical (3) Morillo, E.; Undabeytia, T.; Maqueda, C. EnViron. Sci. Technol. 1997, 31, 3588-3592. (4) Carlisle, S. M.; Trevors, J. T. Water, Air, Soil Pollut. 1988, 39, 409420. (5) Carlisle, S. M.; Trevors, J. T. Water, Air, Soil Pollut. 1986, 29, 189203. (6) O’Neill, P. EnVironmental Chemistry; Chapman & Hall: New York, 1993.
10.1021/ic000849g CCC: $20.00 © 2001 American Chemical Society Published on Web 07/19/2001
Glyphosate Protonation and Cu(II) Complexes of Glyphosate
Figure 1. Theoretical structures of PMG obtained by the HartreeFock theory (basis set 3-21G*). The bond distances (in angstroms) show the distances between atoms potentially capable of intramolecular hydrogen bonding. The four water molecules included in the calculations have been omitted for the sake of clarity.
models are currently being developed as a tool for determining speciation in natural waters. Although copper(II)-PMG complexes have been discussed in the literature previously,7-13 there is an obvious need to provide additional support for proposed structures. The structure of a Cu(II) chelate of PMG as determined by single-crystal X-ray diffraction methods is described by Clarke et al.7 The presence of two chelate rings involving Cu(II) and the amine, carboxylate, and phosphonate groups was demonstrated. Similarly, in solution, the structure of the negatively charged 1:1 complex, CuL(where L represents a PMG ligand), is suggested to consist of two chelate rings in an equatorial plane where the copper(II) ion is complexed through the amine, the carboxylate, and the phosphonate group while a water molecule occupies a fourth position in this equatorial plane.8-11 However, these studies did not provide direct structural information such as bond distances and coordination numbers. There is less agreement regarding the structure of the neutral, protonated 1:1 complex, CuHL. Daniele et al.8 discuss possible (7) Clarke, E. T.; Rudolf, P. R.; Martell, A. E.; Clearfield, A. Inorg. Chim. Acta 1989, 164, 59-63. (8) Daniele, P. G.; De Stefano, C.; Prenesti, E.; Sammartano, S. Talanta 1997, 45, 425-431. (9) Motekaitis, R. J.; Martell, A. E. J. Coord. Chem. 1985, 14, 139-149. (10) McBride, M. B. Soil Sci. Soc. Am. J. 1991, 55, 979-985. (11) Glass, R. L. J. Agric. Food Chem. 1984, 32, 1249-1253. (12) Jezowska-Bojczuk, M.; Kiss, T.; Kozlowski, H.; Decock, P.; Barycki, J. J. Chem. Soc., Dalton Trans. 1994, 6, 811. (13) Buglyo´, P.; Kiss, T.; Dyba, M.; Jezowska-Bojczuk, M.; Kozlowski, H.; Bouhsina, S. Polyhedron 1997, 16, 3447-3454.
Inorganic Chemistry, Vol. 40, No. 17, 2001 4303 structures but favor a structure similar to that of CuL- with all three donor groups involved. Motekaitis and Martell,9 on the other hand, state that only the phosphonate oxygens are involved in chelation within the CuHL complex and that the proton resides on the amine group. McBride10 discusses yet another alternative where the copper(II) ion is coordinated via the amine and either the phosphonate or the carboxylate group to form only one chelate ring. Of these two possibilities, the author considers coordination via carboxylate to be the more probable. The structure of the 1:2 complex, CuL24-, is less well understood. Spectrophotometric measurements by Daniele et al.8 showed no evidence of two nitrogen donors in the equatorial plane. The electron spin resonance studies by McBride,10 however, indicated the presence of two amine and probably two carboxylate groups in the equatorial plane. Heineke, Franklin, and Raymond’s study of the coordination chemistry of glyphosate14 (primarily focusing on metal(III) complexes) highlighted the likelihood of an isomeric mixture of CuL24- complexes. As an attempt to solve some of the discrepancies in the literature and to form a sound basis for continued work involving ternary surface complexes, the current paper presents new evidence concerning the protonation steps of PMG and the structures of copper(II) complexes of PMG, CuHL, CuL-, and CuL24-, in aqueous solution. IR and EXAFS spectroscopies are utilized as complementary techniques whereby IR monitors the behavior of the donor groups of PMG, while EXAFS studies the environment surrounding copper(II). This is the first time EXAFS results are presented for this system and previously undetermined bond distances and coordination numbers are reported. The IR analyses play a crucial role in distinguishing between the varying degrees of protonation, both for PMG alone and for the 1:1 Cu(II)-PMG complexes. Experimental Section Preparation of Solutions. Series of PMG and copper-PMG solutions were prepared for examination with IR and EXAFS spectroscopies. For the IR experiments, PMG solutions were prepared both in H2O and in D2O. Speciation diagrams were generated with the program Solgaswater,15 using equilibrium constants found in the literature,16,17 (Table 1a) to determine the pH and mixing ratios that allow each individual complex to dominate in solution, with concentrations of interfering species at a minimum (Figure 2). On the basis of the results of these equilibrium calculations, solutions containing CuHL, CuL-, CuL24-; H3L, H2L-, HL2-, L3-; D3L, D2L-, DL2-, L3-, as the dominant species were prepared. The resulting concentrations and pH (or pD) values are listed in Table 1b. The solutions were prepared at 25 °C from PMG (Sigma) with 95% purity and CuCl2‚2H2O (Aldrich) with >99% purity, and, to the extent possible, with a constant ionic medium of 0.1 M NaCl. A combination glass electrode was used to measure pH. This was externally calibrated by titrating standardized hydrochloric acid in 0.1 M NaCl. For solutions with pH > 7, argon was bubbled through the solutions to avoid carbonates. NaOH was used to adjust samples to the desired pH. Boiled Milli-Q plus 185 water was used in the preparation of all H2O-based solutions. For the D2O-based solutions, solid H3L was first dissolved in ∼3 M NaOH and water was removed by boiling, followed by drying overnight at ∼90 °C. The resulting solid was then dissolved in D2O (99.9 atom % D; Aldrich), and DCl (diluted from 35% DCl in D2O; Aldrich) was added to reach the required pD. (14) Heineke, D.; Franklin, S. J.; Raymond, K. N. Inorg. Chem. 1994, 33, 2413-2421. (15) Eriksson, G. Anal. Chim. Acta 1979, 112, 375-383. (16) Smith, R. M.; Martell, A. E. Critical Stability Constants; Plenum Press: New York, 1989; Vol. 6. (17) Baes, C. F.; Mesmer, R. E. The Hydrolysis of Cations; John Wiley and Sons: New York, 1976.
4304 Inorganic Chemistry, Vol. 40, No. 17, 2001
Sheals et al.
Table 1 (a) Equilibrium Constants from the Literature,16 Used To Determine the pH and Mixing Ratios That Allow Each Individual Complex To Dominate in Solution equilib reactn
log Ka (25 °C) I ) 0.1 M
H+ + L3- a HL2H+ + HL2- a H2LH+ + H2L- a H3L
10.14 5.46 2.20
equilib reactn
log Ka (25 °C) I ) 0.1 M
equilib reactn
log Ka (25 °C) I ) 0.1 M
Cu2+ + L3- a CuLCu2+ + 2L3- a CuL24-
11.93 16.02
CuL- + H+ a CuHL CuOHL2- + H+ a CuL-
3.92 9.87
(b) Solution Concentrations, pH Values, and Relative Amounts of Dominating Complex in All IR and EXAFS Samples IR [L]tot (mM)
[Cu]tot (mM)
EXAFS pH (or pD)
% [L]tot in complex ∼99 ∼99 ∼95 ∼94
in H2O L3HL2H2LH3L in D2O L3DL2D2LD3L
24 24 24 24
12.3 8.2 4.0 1.0
30 30 30 30
13 8 4 1
CuLCuHL CuL24-
24 24 24
a
24 40 12
6.4 2.9 9.6
[L]tot (mM)
[Cu]tot (mM)
pH
% [Cu]tot in complex
36 77 67
30 23 25
5.6 2.6 9.6
∼98 ∼92 ∼97
∼100 ∼99 ∼95 ∼94 ∼100 ∼87 ∼80
The same log K values were used for the deuterated complexes (DL2-, D2L-, D3L) as they are not expected to differ significantly.
Figure 2. Example of Cu(II)-PMG speciation diagram generated with the program Solgaswater: IR sample of CuHL. Fi is the fraction of PMG found in respective solution species. [Cu(II)] ) 40 mM, [PMG] ) 24 mM. IR Spectroscopy. Attenuated total reflectance (ATR) IR spectra were collected using a Perkin-Elmer 2000 FTIR spectrometer fitted with a deuterated triglycine sulfate (DTGS) detector. The spectra were recorded with a horizontal ATR accessory and a diamond crystal as the reflection element (SensIR Technologies). The sample cell was purged with nitrogen gas throughout data collection to exclude carbon dioxide and water vapor. The angle of incidence for the setup is approximately 45°, which is far from the critical angle. This and the fact that the bands analyzed are weak (0.02), and there was no significant improvement in the fit, compared to including only equatorial oxygen/nitrogen atoms. The second shell of neighboring atoms was best fit with three carbon atoms at an average distance of 2.81 Å, and one phosphorus atom at a distance of 3.01 Å. Taking these values (31) Magini, M. J. Chem. Phys. 1981, 74, 2523-2529.
Figure 4. k3-weighted EXAFS spectra (data and fit) and respective Fourier transforms of the data (not corrected for phase shift) for the Cu(II)-PMG complexes.
into account, the most likely structural interpretation is that two chelate rings and a water of hydration are located in the equatorial plane with all three donor ligands involved in chelating and the remaining water of hydration of the first coordination sphere occupying axial positions. The structure is represented in Figure 5 and agrees well with the findings of Daniele et al.8 and Glass.11 The results verify that the fivemembered chelate rings identified in the X-ray crystal structure
4308 Inorganic Chemistry, Vol. 40, No. 17, 2001
Figure 5. Molecular structure of CuL- (not to scale). The dashed lines represent the four single scattering paths, and the dotted line represents the multiple scattering path, identified in the EXAFS data analysis and described in the text. Od represents a distal oxygen atom (i.e., not bonded to the absorbing copper atom).
of the Cu(II) chelate of PMG7 are also intact when Cu and PMG form CuL- in solution. Fitting the EXAFS data for the protonated complex, CuHL, resulted in the same distances and coordination numbers as for CuL-, with only a slight increase in the Debye-Waller factors of the second-shell features and the distal oxygens. This clearly shows that the structure of CuHL is essentially identical to that proposed for CuL-, i.e., two chelate rings and a water molecule occupying an equatorial plane. It is not possible to define a single structure for the 1:2 complex from the EXAFS data, and as expressed earlier, it is believed that the 1:2 complex does indeed form a mixture of isomers in solution. The distances and coordination numbers given therefore represent a weighted average of all the existing isomers, with more dominant isomers providing a greater contribution to the EXAFS signal. Although EXAFS data analysis using the same coordination numbers as in the 1:1 complexes and distances similar to those of the 1:1 complexes provided a reasonable fit, a significant improvement in the fit was obtained by systematically varying the coordination numbers in both the first and second shells (see different fit results listed in Table 4). A decrease in magnitude compared to the 1:1 complexes is found in the first transition of the FT (Figure 4). This is to be expected for a mixture of isomers. Different arrangements of the donor ligands around the central Cu atom would cause slight differences in the Cu-O bond lengths, depending on the geometry of the complex, i.e., whether the ligands are equatorially bound, axially bound, or a combination of both. These slightly out of phase contributions to the overall signal would to some extent cancel each other out, leading to a decrease in magnitude of the signal and an apparent decrease in coordination number. The best fit for the first shell gave a coordination number of 3.5. Increasing the number of carbon backscatterers in the second shell also produced a better fit to the data. This is conceivably due to the presence of two ligands coordinated in some way to the absorbing copper atom. Note that the coordination number of 1 for phosphorus (Table 4) in this 1:2 complex is the same value as for the 1:1 complex. This implies that if two phosphonate donors are bound to the copper, one is bound at an equatorial position, with low enough thermal disorder to contribute significantly to the EXAFS signal, while the other is bound at an axial position, forming a weaker bond. This axial phosphonate group experiences relatively more disorder, enough that it does not seemingly contribute to the EXAFS signal. There are therefore indications that more dominant isomers have PMG ligands bound to copper(II), via
Sheals et al. equatorial and axial positions, with both phosphonate and carboxylate donor groups responsible for chelation at the axial positions. The slight tendency of the bond distances, R, to increase in the 1:2 complex compared with the 1:1 complexes, for all the paths identified, is likely to be the steric effect of an increased number of neighbors accommodated around the central copper ion. The existence of chelate rings in the equatorial plane of the structures is strongly supported by the third shell observed in the FT of the data (Figure 4), which is interpreted in a similar manner for all three complexes. The distal oxygen atoms (those not coordinated to the copper atom) of the phosphonate group and the carboxylate group provided a good fit to the data by including both contributions from single scattering paths, CuOd (Figure 5), and that of a multiple scattering path through the carbon atom (Figure 5). According to the FEFF calculations, these particular paths were predicted to have relatively large amplitude ratios (44% and 40%, respectively, for a coordination number of 2, while the amplitude ratio of a first-shell oxygen is predicted to be 90-100%). The contributions can be justified by the rigidity of the structure provided by the chelate ring, together with a Cu-C-O bond angle that is sufficiently linear for multiple scattering effects to become significant. Fitts et al.32 describe similar contributions from a third shell in glutamatecopper(II) complexes. The single scattering path was best fit with a coordination number of 2, and this is believed to represent the distal oxygens of the phosphonate group. Although the distal oxygen of the carboxylate could provide an alternative explanation, the average distance of 3.71-3.73 Å obtained in the final fit does not agree well with the 4.06-4.11 Å multiple scattering distance of this oxygen. Summary Different steps of PMG protonation have been identified and structures of Cu(II)-PMG complexes in aqueous solution have been characterized with IR and EXAFS spectroscopic measurements and ab initio frequency calculations. The IR results and ab initio calculations indicate that protonation of L3- forms a HL2- zwitterion whereby the carboxylate and phosphonate donor groups are intramolecularly hydrogen-bonded to the amine group, R2NH2+. The zwitterion remains intact on progressing to H2L-, and the second proton resides on the phosphonate group. For H3L, the structure is still zwitterionic but intramolecular hydrogen bonding is predominantly between the phosphonate and amine groups only and a COOH group is formed. Structures of CuL-, CuHL, and CuL24- contain fivemembered chelate rings with the amine, carboxylate, and phosphonate groups of the PMG ligand all involved in chelation. For CuL-, evidence of monodentate linkage to carboxylate and a direct bond to the phosphonate were identified with IR and there was evidence for H2O molecules in the first coordination sphere. The EXAFS data confirmed these findings, indicating two chelate rings oriented in the equatorial plane, with all three donor groups of the PMG molecule responsible for chelation. CuHL was found to be essentially identical to CuL-, differing only in the protonation of the phosphonate group, as identified by IR. Although a single definitive structure of the CuL24complex could not be identified, there is substantial evidence for the existence of several isomers. IR spectral features for CuL24- are similar to those for CuL-, but peaks in the (32) Fitts, J. P.; Persson, P.; Brown, G. E., Jr.; Parks, G. A. J. Colloid Interface Sci. 1999, 220, 133-147.
Glyphosate Protonation and Cu(II) Complexes of Glyphosate
Inorganic Chemistry, Vol. 40, No. 17, 2001 4309
carboxylate region suggest that at least some carboxylate groups must be chelated at axial positions. The absence of a peak representing H2O molecules in the first coordination sphere indicates the chelation of PMG donors at both equatorial and axial positions while the phosphonate peaks suggest the coordinated PMG to be present in a range of chemical states. EXAFS results also point toward an isomeric mixture of complexes, primarily indicated by a loss in magnitude in the first transition of the Fourier transformed EXAFS data. These results form the basis for forthcoming studies of ternary surface complexes formed from adsorption of Cu(II) and PMG onto mineral oxide surfaces.
Dr. John Bargar, for their help and advice. SSRL is operated by the Department of Energy, Office of Basic Energy Sciences. We also acknowledge the National Institutes of Health, National Center for Research Resources, Biomedical Technology Program, and the Department of Energy Office of Biological and Environmental Research (OBER), which support the SSRL Structural Molecular Biology Program, whose instrumentation was used for the measurements. The Kempe Foundation, Sweden, and the OBER are acknowledged for providing financial support for the educational stay of Julia Sheals at SSRL. Professor Staffan Sjo¨berg is acknowledged for his valuable comments on the manuscript and Dr. Magnus Karlsson for his assistance with the illustrations. Mrs. Agneta Nordin is gratefully acknowledged for help with the IR measurements.
Acknowledgment. This work was supported by the Swedish Natural Science Research Council. We thank the staff of Stanford Synchrotron Radiation Laboratory (SSRL), particularly
IC000849G